Project 2 of the CREATE Pharmacogenetic Research Network will determine the genotype-gene expression relationship in therapeutically relevant tissue. This project will use human tissue to validate the relevance of identify polymorphisms, rather than in vitro or surrogate tissue approaches that often are not borne out during clinical application. Therefore, Project 2 will address the following Specific Aims. 1. Determine the genotype-gene expression relationship in malignant tissue to validate gene variants 2. Establish the nomenclature for mutations in the pathway genes. 3. Evaluate statistical genetics approaches for assessment of multiple variants in a functional haplotype Patient genotype of all known variants and those identified from Project 1 will be determined in 1000 GI tumor specimens and adjacent normal tissue. Gene expression arrays will be used to provide comprehensive information on all of the genes of interest. Quantitative assays of protein and functional activity will also be employed in these tissues. This will provide a direct validation of the functional impact of defined alleles in therapeutically relevant tissue. One of the most important statistical challenges of pharmacogenetics is the ability to formally analyze and organize enormous amounts of polymorphism data become available. Statistical genetics analysis will also be conducted to develop and test approaches for validating the functional impact of single mutations in the context of multiple variants within a gene and for the assessment of multiple comparisons of genotype on a clinical or biological phenotype.